The Agent Skills Directory

[COMMAND_EXECUTION]: The statusline-command.sh script uses eval to process JSON input from the agent environment. Although it employs jq's @sh filter for escaping, evaluating strings derived from potentially attacker-controlled sources (such as workspace directory names or git worktree names) is a risky practice that could lead to command injection if escaping is bypassed.
[COMMAND_EXECUTION]: The skill's installation process requires modifying the agent's persistent configuration file (~/.claude/settings.json) to execute the status line script. This creates a persistence mechanism where a local script is automatically executed by the agent for every interaction.
[COMMAND_EXECUTION]: The installation steps include changing file permissions via chmod +x on the status line script to facilitate its execution.
[DATA_EXFILTRATION]: The skill accesses the agent's private configuration file (~/.claude/settings.json) to read and merge settings. Access to such files exposes internal agent configuration and operational metadata.

statusline